#74068 nth child specificity #369
Conversation
src/services/selectorPrinting.ts
Outdated
| // Test for presence of complex-selector-list S; find highest specificity | ||
| const keyword = text.indexOf(' of '); | ||
| if (keyword !== -1) { | ||
| const psuedoSelector = element.getChildren(); |
There was a problem hiding this comment.
Typo - should be named 'pseudoSelector' (e/u inversion)
src/services/selectorPrinting.ts
Outdated
| const selectorText = text.slice(text.indexOf('(') + 1, text.length -1); | ||
|
|
||
| // Test for presence of complex-selector-list S; find highest specificity | ||
| const keyword = text.indexOf(' of '); |
There was a problem hiding this comment.
Testing for a space before and after of is not clean. There could also be a tab, or even a non word character
ul.one > li:nth-child(-n + 3 of.noted) {
}
would be valid too.
I suggest to jeust use the scanner to find out if there's an of ident in the string.
There was a problem hiding this comment.
Reworked the logic in the latest commit. When the 'An + B [of S]?' syntax is respected then we can just grab the last childElement as it will contain the entire complex-selector list (S).
The edge case where 'n' has no integer prefix (the parser does not recognise it as a binary expression token). Extract out the complex-selector list, and re-parse the selectors. This also handles the different syntax possibilities; added more tests for completeness.
fix(cssParser): parseSelector; optional boolean test(): additional test for complex-selector list syntax
|
Looks good! If you can run the fomatter on |
Amended |
|
Thanks a lot! |
Fix issue #74068 "[css] Wrong selector specificity values with :nth-child(even) and :nth-child(odd)". This patches the specificity for the nth-child/nth-last-child psuedo-selector.
Debugging the code it seems the scanner/parser treats words 'even', 'odd', and 'n', as NodeType 16 (ElementNameSelector), which contributes to a false specificity.
'div:nth-child(n of li)' should produce a specificity of 0-1-2, but produces 0-1-4 suggesting all three arguments are treated as elements. If the 'n' has an integer prefix the correct specificity is produced however.
Rather than try to unpack the scanner/parser logic, the solution exploits the formula of the psuedo-selector by testing for the presence of " of " within the arguments string, "An+B [of S]?".